# Inworld Realtime TTS 2 - Core.Today AI API

> Inworld's most expressive TTS with natural-language steering — place bracketed instructions like [speak quickly] before the text they apply to. Real-time latency and 15+ language support.

- **Provider**: Inworld
- **Model ID**: inworld/realtime-tts-2
- **Category**: Audio & TTS
- **Credits**: 59 per 1000 characters (0.057 credits/char, billed per character)
- **Speed**: Fast
- **Quality**: High

## Features
- Natural-language steering with bracketed instructions
- Real-time latency
- 15+ language support with auto detection
- Preset voices or custom cloned voice IDs
- Text normalization control for numbers and dates

## Use Cases
- Interactive voice agents with directed delivery
- Real-time character voices for games
- Dynamic IVR and announcement systems
- Expressive chatbot voices
- Live content narration

## API Endpoint

Base URL: https://api.core.today/v1
Create Prediction: POST /predictions
Get Status: GET /predictions/{job_id}

## Authentication
Header: X-API-Key: YOUR_API_KEY

## Input Parameters

### Required
- **text**: string - Text to convert (max 2,000 chars). Supports bracketed steering like [speak quickly] before the text it applies to

### Optional
- **voice_id**: string (default: Ashley) - Preset voice name (Ashley, Dennis, Alex, Darlene) or a custom cloned voice ID
- **language**: string (default: auto) - Language of the input text. 'auto' detects automatically
  Options: auto, en, zh, ja, ko, ru, it, es, pt, fr, de, pl, nl, hi, he, ar
- **speaking_rate**: number (default: 0) - Speaking speed multiplier (0-1.5). 0 = normal speed (1.0)
- **temperature**: number (default: 0) - Randomness control (0-2). 0 uses the model default
- **audio_format**: string (default: mp3) - Output audio format
  Options: mp3, wav, ogg_opus, flac
- **sample_rate**: integer (default: 48000) - Audio sample rate in Hz
  Options: 8000, 16000, 22050, 24000, 32000, 44100, 48000
- **text_normalization**: string (default: auto) - Expand numbers/dates/abbreviations before synthesis
  Options: auto, on, off

## Examples

### Steered Delivery
Direct the speaking style with a bracketed instruction
```json
{
  "model": "inworld/realtime-tts-2",
  "input": {
    "text": "[speak quickly with a clear and direct manner] Your confirmation number is 8 4 7 2 9. Your order will arrive by FRIDAY at 3:45 PM.",
    "voice_id": "Dennis",
    "audio_format": "mp3"
  }
}
```

## Response Format
```json
{
  "job_id": "abc123",
  "status": "pending | processing | completed | failed",
  "result": "URL or data (when completed)"
}
```

## Usage Flow
1. POST /predictions with model and input -> receive job_id
2. GET /predictions/{job_id} -> poll until status is completed or failed
3. Result contains output URL(s)

## Tips
- Place steering instructions like [speak quickly] or [whisper softly] right before the text they apply to
- Use language: auto for mixed-language text, or pin the language for consistent pronunciation
- Turn text_normalization on for confirmation numbers, prices, and dates
- Keep each request under 2,000 characters

## Documentation
https://replicate.com/inworld/realtime-tts-2